Custom AI Tool
A small tool costs $9,500 to build, then $450 a month, or no monthly fee if your team takes it over. We build one AI tool for a job your team repeats every week, such as checking files, answering staff questions from your policies or comparing payer rates. A named person approves every result.
Prefer to skip the call? Get a written proposal
Built for
- A task that takes someone most of their week and isn't covered by our other AI tools
- A prototype from an AI Readiness Sprint that needs to become a working tool
- Teams with engineers who want the code, prompts and runbook handed over
- Health-tech sales teams that want a weekly note, from public sources, when a target account changes leaders, budgets or vendors
Why it's worth it
Today
Someone on your team spends most of the week on a repeatable job: reading files, checking them against a list, retyping figures from one system into another. At 30 hours a week, that's about $34,000 a year of a medical secretary's time, and the cases that need judgment wait behind the routine ones.
After
The tool does the repeatable part and queues each result for that person to approve, so their hours go to the cases that need judgment. It runs in your accounts, and you own it.
Why it matters
$34,000
Thirty hours a week of a medical secretary's time costs about $34,000 a year at the median wage of $22.08 an hour in May 2025 (30 hours × 52 weeks × $22.08).
How it works
A person on your team approves every result
The tool runs these steps on a schedule. Nothing moves past the approval step until someone on your team signs off.
- InputYour workflow, mappedWe map the job with the people who do it, then the tool reads the information from where it is kept: document folders, spreadsheets, a vendor portal or summary exports.
- AI modelRead, check or draftThe tool does the repeatable part, such as reading documents, checking them against your rules or drafting replies.
- A person approvesYour reviewer decidesEvery result waits for a named person, and unclear cases go straight to them.
- OutputDone where you workResults go back into the software your team already uses for that work, or out by email, Teams or a shared folder, with a log of who approved what.
What you get
Credentialing
File audit and tracker
Each clinician's file checked for missing documents and expiration dates, including every state license a virtual care clinician holds. The tracker is a small tool: $9,500 to build, then $450 a month. Your coordinator still submits applications and calls payers.
Staff questions
A policy assistant
Answers staff questions from your own policy manual, in Teams or on your intranet, and cites the page. If your organization already licenses an AI assistant that can search your documents, try that first.
Payer rates
Fee schedule comparison
Your fee schedule set against what each payer paid, by code, from summary exports. It shows where each payer pays least, in time for a contract renewal.
Anything else
Your workflow, simplified first
Every step and handoff mapped with the people who do the work, and steps that add nothing removed before anything is built.
What done means
The tool passes a set of test cases we agree on, at an accuracy we agree on, then runs for five business days in a row without help, and your team gets recorded training.
The code, prompts and runbook are handed over, so your team could run the tool without us.
Choose your version
Each version is a one-time build, then a flat monthly fee for monitoring and fixes, or a handoff to your own team. Either way, a tool that reads your files keeps running in your account.
What the monthly fee covers
We watch every run, fix the tool when a source or system changes, keep it working as AI models change, and reply by the next business day, with a check-in each month. It's $450 a month for a small tool and $450 to $1,500 for a full build, by size, month to month with 30 days' notice. Or skip it: we hand your team the code, prompts and runbook, and the tool keeps running in your accounts. A tool that reads your files runs in your own Microsoft 365, Google Workspace or cloud account, so you pay its hosting and AI usage directly, and your proposal estimates that bill in writing before you sign. We set the accounts up with you during the build, and their bills come to you. One that reads only public sources can run on our servers, with those costs included.
Small tool
For one well-defined job
Build price: $9,500
+ $450 a month
About 2–3 weeks of build work
One job, such as checking files or answering policy questions, with one person approving the results.
What's included in Small tool
- A test set built from real examples of the work in your organization
- The approval queue and a log of who approved what
- 30 days of fixes after launch
- Or skip the monthly fee and take the tool over yourselves
Year one, build plus 12 months of our fee$14,900
Plus hosting and AI usage in your own accounts, estimated in your proposal before you sign.
Full build
For work across several systems
Build price: from$25,000
Up to $60,000, depending on how many systems it connects to.
+ $450 to $1,500 a month
About 5–12 weeks of build work
Several steps or systems, mapped and simplified before anything is built.
What's included in Full build
- Workflow map and simplification first
- Tested on real examples from your own organization, at an agreed accuracy
- Code, prompts and a written operating guide (a runbook) handed over
- Monthly fee set by size, or none if your team takes it over
Year one, build plus 12 months of our fee$30,400–78,000
Plus hosting and AI usage in your own accounts, estimated in your proposal before you sign.
Is the job credentialing? Start with an audit of your credentialing files for $1,500 (up to 8 clinicians) or $3,500 (up to 25), one time, credited toward a tracker within 90 days.
Payment and terms
- Under $20,000, half at signing and half at delivery. From $20,000, 40% at signing, 30% at a mid-build demo and 30% at acceptance.
- Paying by purchase order? There's no deposit: we invoice half at kickoff and half at acceptance, or from $20,000, 25% at kickoff, 35% at a mid-build demo and 40% at acceptance. Acceptance means the “What done means” criteria are met.
- The monthly fee is month to month with 30 days' notice, and we reply by the next business day. The tool is yours on final payment and keeps running in your account if you stop the fee.
- The fee for a smaller first step counts toward the full build if you buy it within 90 days.
- Running three or more of our AI tools? Their monthly fees are capped at $1,000 a month in total. Health system, network and multi-site versions are priced on their own.
- Anything outside the written scope is priced as a change order before it starts.
Year one is the build plus 12 months of our fee. Hosting and AI usage are billed to you by those providers, and your proposal estimates them before you sign. The build times on this page are measured in weeks of full-time work. Start and finish dates are set in writing in your proposal.
Tyler Kent does this work himself. He has 13+ years in data science and analytics, most of it in healthcare. About Tyler
Where this comes from
Every system on Tyler's Portfolio page follows the same rule. Tessera, Emblema, Sentinel and Opus, four AI systems he built, each do one job, record what they did, and leave the decision to a person. A custom tool is built the same way, around your workflow.
See the four systemsQuestions
Small tool or AI Readiness Sprint?
If you know the job, buy the small tool. If you don't know which job to start with, the AI Readiness Sprint finds it in two weeks, and its fee counts toward the build within 90 days.
Will it make clinical decisions?
No. Custom AI tools don't diagnose, triage or recommend treatment. Statistical risk models, such as which patients to call first after discharge, are a separate service, Predictive Models, tested on your own history and checked for bias. Those that need patient data start once a BAA and our insurance are in place.
How is this different from the other AI tools on this menu?
Each of those does a set job at a set price. This is for a job that isn't on the list. Its scope and price are fixed in writing before the build starts.
Can you build tools that use patient data?
Not yet. Builds that read patient records, such as a practice back-office inbox or a home health referral desk, start once a BAA and our insurance are in place. To hear when they open, say so when you ask for a written proposal. If the problem is patients waiting for beds, start with the Throughput Diagnostic.
When is a custom app the better choice?
When your team needs a screen to work in all day, such as a checklist, a tracker or a schedule. A custom AI tool does a job for the team and asks a person to approve the result.
Who owns it?
You own your data and every output, and the tool is yours on final payment. Parts we reuse from project to project are licensed to you permanently for your own use. A tool that reads your files already runs in your accounts, so it keeps running if you stop support.
How an engagement worksQuestions about prices, payment and patient data
Book a 20-minute call
Tell us which task eats the most hours. If we can help, you'll get a fixed-price proposal within 48 hours.
Prefer to skip the call? Get a written proposal
Or email tyler@kentapplied.com, or connect on LinkedIn.